ARTICLE

Received 26 Oct 2015 | Accepted 2 May 2016 | Published 13 Jun 2016 DOI: 10.1038/ncomms11804 OPEN Efficient generation of monoclonal antibodies against peptide in the context of MHCII using magnetic enrichment

Justin A. Spanier1,*, Daniel R. Frederick2,*, Justin J. Taylor3,*,w, James R. Heffernan1, Dmitri I. Kotov3, Tijana Martinov1, Kevin C. Osum1, Jenna L. Ruggiero1, Blake J. Rust4, Samuel J. Landry4, Marc K. Jenkins3, James B. McLachlan2,* & Brian T. Fife1,*

Monoclonal antibodies specific for foreign antigens, auto-antigens, allogeneic antigens and tumour neo-antigens in the context of major histocompatibility complex II (MHCII) are highly desirable as novel immunotherapeutics. However, there is no standard protocol for the efficient generation of monoclonal antibodies that recognize peptide in the context of MHCII, and only a limited number of such reagents exist. In this report, we describe an approach for the generation and screening of monoclonal antibodies specific for peptide bound to MHCII. This approach exploits the use of recombinant peptide:MHC monomers as immunogens, and subsequently relies on multimers to pre-screen and magnetically enrich the responding antigen-specific B cells before fusion and validation, thus saving significant time and reagents. Using this method, we have generated two antibodies enabling us to interrogate antigen presentation and T-cell activation. This methodology sets the standard to generate monoclonal antibodies against the peptide–MHCII complexes.

1 Department of Medicine, Center for Immunology, University of Minnesota Medical School, Minneapolis, Minnesota 55455, USA. 2 Department of Microbiology and Immunology, Tulane University School of Medicine, New Orleans, Louisiana 70112, USA. 3 Department of Microbiology and Immunology, Center for Immunology, University of Minnesota Medical School, Minneapolis, Minnesota 55455, USA. 4 Department of Biochemistry, Tulane University School of Medicine, New Orleans, Louisiana 70112, USA. * These authors contributed equally to this work. w Present address: Vaccine and Infectious Disease Division, Fred Hutchinson Cancer Research Center, Seattle, Washington 98019, USA. Correspondence and requests for materials should be addressed to B.T.F. (email: bfi[email protected]).

NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804

he general approach for generation of monoclonal (APC) using SA-PE-AF647 or SA-APC-DyLight 755, compared antibodies (MAb) reactive to a defined protein antigen to a decoy p:MHCII-APC reagent (Fig. 2a). Three distinct subsets Thas been well documented since the original report in of antigen-specific B cells (p:MHCII specific, MHCII specific and 1975 by Drs. Kohler and Milstein1. The utility and broad use decoy p:MHCII specific)10,20 were evaluated for GL7 and of MAbs in biological systems earned Kohler and Milstein the intracellular Ig expression associated with mature germinal Nobel Prize for medicine in 1984 (ref. 2). In this report we center B cells. Phenotypic analysis demonstrates the p:MHCII- describe a novel methodology to specifically and reliably generate PE þ population is enriched for mature germinal center B cells MAbs that target peptide in the context of MHCII, which has (GL7 þ and intracellular Ig À ) demonstrating successful priming only occurred a few times since 1975 (refs 3–9). and T-cell help (Fig. 2a). We verified the enrichment approach at To generate a MAb using the traditional approach, mice are day 3 post antigen boost, before hybridoma fusion. Magnetic immunized, the responding B cells are isolated, fused to myeloma enrichment resulted in an increase to 11.1% of the B cells staining cells with hypoxanthine–aminopterin–thymidine (HAT)-based positive for p63:IAg7-PE tetramer, and phenotypic markers selection, screened and sub-cloned to isolate monoclonal demonstrating the presence of germinal center B cells within hybridomas2. Screening requires the examination of hundreds this population (Fig. 2b). The enriched fraction contained or even thousands of clones for one MAb, creating a major 2.1 Â 107 cells, which was 23-fold reduced compared with the bottleneck. This approach typically yields o1–5% hybridomas starting population. These cells were subsequently fused with specific for a protein target antigen causing a prominent hurdle, SP2/0 myeloma cells and plated onto ten, 100 mm plates both in time and resources4. However, this method is not containing semi-solid media under HAT selection. Fourteen specifically designed to generate peptide:MHCII (p:MHCII) days after plating, 190 colonies were picked and screened by reactive MAbs, and B-cell tolerance against self MHC adds enzyme-linked immunosorbent assay (ELISA). Without to the difficulty. To overcome this, we developed a enrichment, we would have required 50 plates to screen 5 Â 108 novel methodology to generate MAb against a specific cells. We predict that these 50 plates would have contained at p:MHCII complex. B-cell clones specific for the antigen of least 5,000 colonies, most of which could not have been selected interest are enriched immediately before myeloma fusion, thus or screened for further analysis due to time and reagent significantly reducing the screening required. The basis for this constraints. Thus, enrichment allowed us to screen every visible methodology centers on having a stable p:MHCII monomeric colony, and saved significant time and reagents. protein linked to biotin as the B-cell antigenic target. This approach has several advantages. First, immunization with p:MHCII complexes induces a B-cell response specific for that Screening and in vitro validation of p:MHCII MAb.After peptide in the context of MHCII. Second, use of antigen-specific expansion of each colony, secreted antibody in the culture super- tetramers allows us to pre-screen immunized mice to confirm the natant was assessed for binding to p63:IAg7 compared to decoy g7 expansion of p:MHCII-specific B cells. Third, it offers the ability InsB10–23:IA by ELISA. Thirty-two of the 190 colonies produced 10 g7 to enrich for antigen-specific B cells while discarding B-cell antibodies that bound to both, p63:IA and InsB10–23: clones responding to unrelated antigens. Specifically, the utility IAg7, indicating specificity for an IAg7 epitope (Table 1). In con- of a site-directed protein biotinylation allows for the enrichment trast, 11 hybridomas produced antibodies that bound only of B cells reactive to the target protein/peptide by generating a p63:IAg7 (34.4% success rate for p:MHC or 5.8% overall), sug- tetrameric antigen, thus increasing the avidity of B cells gesting the desired specificity for this peptide bound to IAg7 for antigen and enabling the capture and enrichment of (Table 1). Figure 3a illustrates 20 clones, 10 that reacted to both 10,11 g7 g7 antigen-specific B cells . This results in a significant time p63:IA and InsB10–23:IA (bottom), and 10 that are unique for and cost saving as fewer colonies are required for screening, and a p63:IAg7 (top). We further characterized these 10 clones that higher percentage of selected hybridomas produce MAb uniquely bound p63:IAg7 for TCR blocking ability to limit in vitro against p:MHCII. Finally, this enrichment approach could be antigen-specific T-cell proliferation (Fig. 3b). Splenocytes were used for any MAb protein target including peptides and haptens, isolated from TCR transgenic BDC2.5 mice, labelled with car- not just p:MHCII12–14. boxyfluorescin succinimidyl ester (CFSE) and cultured with p63 peptide in the presence or absence of hybridoma supernatant for 4 days. BDC2.5 splenocytes incubated with peptide alone resulted in Results 87.5% CD4 þ T cells proliferating, while T cells incubated with Generation of p:MHCII MAb. The workflow for this metho- peptide plus hybridoma A1 limited BDC2.5 T-cell proliferation to dology and the necessary steps for p:MHCII MAb generation are 56% (Fig. 3b). The remaining nine hybridomas screened had illustrated in Fig. 1. Generation and validation of p:MHCII MAb varying degrees of inhibition (Fig. 3b). We then used an isotype- can be completed in o8 weeks. We were interested to develop a specific ELISA to determine that A1 antibody was IgG1. A large- reagent to block T-cell receptor (TCR) recognition of a diabetes- scale purification was next performed to obtain purified MAb from relevant peptide14–16. We initially developed antibodies against hybridoma A1 (named FS1). Using the FS1 MAb (anti-p63:IAg7) p63 peptide in the context of IAg7 MHCII molecule, given that we performed an in vitro dose–response assay and demonstrated p63-activated BDC2.5 CD4 þ T cells mediate accelerated 80.5% specific reduction in proliferation with 1.72 mM FS1 MAb autoimmune diabetes when transferred into wild-type non- (Fig. 3c). In contrast, the FS1 MAb only reduced BDC2.5 T-cell obese diabetic (NOD) hosts17–19. We isolated splenocytes from proliferation to another BDC2.5 mimetope (p31) by 5.6% com- five p:MHCII (p63:IAg7) immunized BALB/c mice and pared with an isotype control (Fig. 3c). p63-activated BDC2.5 T magnetically enriched for antigen-specific B cells using PE- cells demonstrated 99.85% reduction in IFNg production when conjugated p63:IAg7 tetramers followed by anti-PE magnetic cultured with 1.72 mM of FS1 MAb, compared with an isotype beads10. To validate successful priming and expansion, we control (Fig. 3c). Importantly, IFNg production by p31-activated analyzed the phenotype of p:MHCII-specific B cells in naive BDC2.5 T cells was not altered (Fig. 3c). We noted a similar trend mice compared to day 7 post immunization (Fig. 2a). Antigen- with IL-17A (Fig. 3c). Taken together, these findings illustrate the specific B cells were identified from immunized mice by specificity of the FS1 MAb as p31 differs from p63 by two amino p:MHCII-PE tetramer excluding those that bound to acids at positions P-1 and P1 of the MHCII binding pocket21.As streptavidin (SA)-phycoerythrin (PE) or SA-allophycocyanin an extension of specific binding, splenocytes from NOD mice were

2 NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804 ARTICLE

Immunize p:MHCII monomer in CFA

Day 7 Stain with target p:MHCII PE and decoy p:MHCII APC

Magnetic enrichment FACS phenotype analysis 14–28 days

Immunization boost p:MHCII monomer i.v.

3 days

Harvest and prepare spleen from immunized host Save aliquot of pre-enriched B cells for 1 h FACS phenotyping Enrichment with anti-PE magnetic beads Save aliquot of post-enriched B cells for 2 h FACS phenotyping Fuse myeloma partner with enriched B cells

24 h

Plate in cloning/selection containing methylcellulose

After 10–14 days

Harvest clones, Transfer to 96 well plates

After 3 days

Expand and screen supernatant by ELISA

6 h

Expand and freeze positive MAb clones

1–5 days

Validate MAb in vitro and in vivo, determine Ab isotype MAb validation phase MAb generation phase Figure 1 | Workflow for the generation and validation of p:MHCII MAb. Mice were immunized with p:MHCII emulsified in complete Freund’s adjuvant. After 7 days splenocytes from naive and immunized mice were collected and stained for specific immunogen p:MHCII-PE and decoy p:MHCII APC, streptavidin-PE-AF647, streptavidin-APC-DyLight 755, magnetically enriched using anti-PE and anti-APC magnetic beads and analyzed by flow cytometry20. P:MHCII-specific B cells were gated from streptavidin, APC and PE binding cells20. Germinal center B cells (GL7 þ and Intracellular Ig À ) and plasma cells (GL7 À and intracellular Ig þ ) were then identified from the various B-cell populations binding these distinct tetramer reagents to demonstrate successful priming. After 28 days, mice were boosted with a second immunization of p63:IAg7 monomeric protein intravenously. Three days following the immune boost, we preformed magnetic B-cell enrichment for splenic B cells binding to the p63:IAg7 tetramer:PE reagent as described in materials and methods. Enriched cells are fused with myeloma fusion partners, expanded and screened for in vitro and in vivo validation. p63 peptide-pulsed and stained with labelled FS1 MAb illustrating significantly increased FS1 MAb staining in response to p63 CD8a þ cDCs (dendritic cells) and B220 þ B cells stained positive peptide-pulsed compared to OVA peptide (P ¼ 0.005 and for p63 peptide but not ovalbumin peptide (OVA141–160)control P ¼ 0.008, respectively), while T cells showed no specific staining (Fig. 3d). The uniform histogram shift suggests a large portion of (Fig. 3e). the DCs and B cells stained with varying levels of FS1 MAb Using this novel methodology, we also generated an antibody demonstrating peptide presentation in vitro (Fig. 3d). Importantly, specific for the peptide 2W12,22 bound to IAb (named W6). Using CD4 þ and CD8 þ T cells did not stain positive for the FS1 MAb this reagent, we validated in vitro antigen loading and (Fig. 3d). Immunostaining was next performed to demonstrate presentation using bone marrow-derived dendritic cells that peptide binding to MHCII in vivo. NOD mice were injected with were pulsed with green fluorescent protein (GFP)-linked 2W p63 or OVA peptide in the footpad and 1.5 h later popliteal lymph peptide. Results in Figure 4a demonstrate, 47% of the node cells were stained with FS1 antibody to identify p63-loaded GFP-positive cells were W6 MAb (anti-2W:IAb) positive and antigen-presenting cells (Fig. 3e). Both DCs and B cells had were mostly CD11c þ CD11b þ double-positive cells. We next

NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804

a 105 2.6 Decoy 4 p:MHCII- 10 specific 3 Naive Immunized 10 5 2 10 0.36 10 0.98 2.1 13 0 0.29

104 5 MHCII- 10 50 specific 4 103 10 0.89 6.1 103

tetramer-APC 2

Decoy p:MHCII 10 102 0 0 12 p:MHCII- 0102 103 104 105 0102 103 104 105 specific 105 66 p:MHCII tetramer-PE 104 103 102 7.1 GL7 0 0102 103 104 105 b Intracellular Ig p:MHCII tetramer+ Enriched Enriched 5 10 105 18.3 104 104 103 103 102 0 102 2 11.1 3.9 –10 0

Flowthrough Flowthrough

105 105 0.99 104 104 103 103 102 0 102 0.36 –102 0.14 B220 GL7 0 0102 103 104 105 –103 0103 104 105 p63:IAg7 tetramer-PE Intracellular Ig

Figure 2 | Enrichment and phenotypic analysis of p:MHCII-specific B cells. (a) Flow cytometric analysis of p:MHCII-specific B cells before and 7 days post immunization with p:MHCII monomer. Splenocytes from naive and immunized mice were collected and stained for specific immunogen p:MHCII-PE and decoy p:MHCII APC tetramers and magnetically enriched using anti-PE and anti-APC magnetic beads. Germinal center B cells (GL7 þ and Intracellular Ig À ) and plasma cells (GL7 À and Intracellular Ig þ ) were then identified from the various p:MHCII-specific B-cell populations binding these distinct tetramer reagents. P:MHCII-specific B cells were gated from streptavidin, APC and PE binding cells using SA-APC-DyLight 755 or SA-PE-AF647 20. (b) Representative flow cytometric analysis of p63:IAg7-enriched antigen-specific B cells obtained before myeloma fusion. Germinal center B cells (GL7 þ and intracellular Ig À ) and plasma cells (GL7 À and intracellular Ig þ ) were then identified within the p63:IAg7-PE tetramer specific B cells. Data are representative of two independent experiments with 2–5 mice per group.

Table 1 | Efficiency of generating hybridomas producing p:MHCII-specific antibodies.

Antigen # Mice immunized # Hybridomas screened # Hybridomas þ for MHCII (%)* # Hybridomas þ for peptide:MHCII (%) p63:IAg7 5 190 32 (16.8) 11 (5.8)w 2W:IAb 2 576 234 (40.6) 5 (0.9)w LLO:IAb 2 576 236 (40.9) 32 (13.5)* Ins:IAg7 2 576 22 (3.8) 11 (1.9)* p31:IAg7 2 480 21 (4.4) 14 (2.9)* mimetope:IAg7 2 480 21 (4.4) 12 (2.5)*

*Specificity as defined by positive ELISA plate compared to decoy pMHCII complex. wSpecificity as defined by functional assay blocking T-cell proliferation or MAb binding to APCs.

4 NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804 ARTICLE

ab 100 p63:IAg7 100 Ins:IAg7 80 80 60 40 60 % of max 20 0 40 F2 F7 A1 A3 A5 B3 C5 C6 C8 G8 100 20

80 % Divided (diluted CFSE) 0 60 p63 peptide –+ + +++++++++ 40 Hybridoma ––A1 C5 C6 F2 G8 B3 A5 A3 F7 C8 % of max 20 0 F5 F8 F9 E5 C7 D8 F11 C6b C11 G11 d Hybridoma supernatant + CD8α cDCs B220+ B cells c No peptide 100 OVA141–160 80 p63

60

40 p63 0102 103 104 105 0102 103 104 105 20 p31 CD8+ T cells CD4+ T cells # of cells

% Divided (diluted CFSE) 0

100,000

) 10,000 –1 0102 103 104 105 0102 103 104 105 1,000 FS1 MAb (pg ml γ p63 e IFN 100 p31 DCs B cells T cells OVA141–160 P63 10

10,000 # of cells ) 3 4 5 3 4 5 3 4 5 –1 1,000 01010 10 01010 10 01010 10 FS1 MAb 100 DCs B cells p63 100 P=0.005 150 P=0.008 10 IL-17A (pg ml p31 80 100 1 60 40 172 50

17.2 1.72 FS1 gMFI 1,720 No Ab 20 FS1 gMFI

No peptide Isotype Ctrl 0 0 OVA p63 OVA p63 FS1 (anti-p63:IAg7) MAb (nM) 141–160 141–160

Figure 3 | Screening and functional validation of the FS1 MAb. (a) ELISA results from twenty independent hybridomas presented as per cent maximum absorbance compared to anti-IAg7 mouse hybridoma clone 10–2.16 as positive control. Plates were coated with p63:IAg7 monomer and compared to g7 InsB9–23:IA monomer-coated plates. Supernatant was added and secondary antibody was used to measure binding by ELISA. Media alone was used for a negative control (dashed line). (b) CFSE proliferation assay in response to p63 peptide was performed using splenocytes from BDC2.5 mice. The per cent of divided CD4 þ T cells is shown with maximum division with peptide alone (87.5%) and inhibition with each hybridoma supernatant screened. Data are representative from two independent experiments. (c) CFSE-labelled BDC2.5 T cells were cultured for 4 days in the presence of p63 or p31 peptide with varied concentrations of FS1 MAb or isotype control IgG1 antibody (1.72 mM). The per cent divided CD4 þ T cells is shown for each concentration of blocking FS1 MAb compared with maximum proliferation with no antibody (no Ab) or isotype Ab. FS1 MAb effects on IFNg cytokine production from the cultured cells (middle panel) and IL-17A cytokine production (bottom panel). Data are representative from two independent experiments in duplicate for each antigen concentration. (d) In vitro antibody staining on antigen-presenting cells following peptide pulse with p63 or OVA141–160 using purified clone FS1 MAb to detect p63 loaded cDCs (CD8a þ CD11c þ MHCII þ , CD3e À , F4/80 À ), and B cells (B220 þ ,MHCII þ ,CD11c À ,CD3e À ,F4/80 À ) but not CD4 þ or CD8 þ T cells (CD3e þ , CD11c À , CD11b À , B220 À , F4/80 À ) compared with no p63 peptide negative control. Data are representative from three independent experiments. (e) In vivo staining of antigen-presenting cells with FS1 MAb following footpad immunization. NOD mice received p63 or

OVA141–160 peptide and 1.5 h following injection the popliteal lymph node was collected and stained for antigen-specific presentation using biotinylated FS1 MAb. FS1 MAb staining was detected on DCs and B cells, but not Tcells using fluorochrome-linked streptavidin. Statistical significance was calculated using a two-tailed Student’s t-test. Data are representative from two independent experiments with 2–4 mice per group. validated in vivo antigen loading and presentation. C57BL/6 mice (Fig. 4b). In a separate in vivo model, we used the W6 MAb to were immunized intradermally with either ovalbumin protein or identify antigen-presenting cells immunized with 2W peptide and 2W-GFP. At 24 h post injection, MHCII þ antigen-presenting two different adjuvants. C57BL/6 mice were immunized with cells from the draining lymph node had increased W6 MAb 2W-GFP and either 50-cytosine-phosphate-guanine-30 (CpG) or reactive populations (15%) compared with 1% of controls double-mutant heat-labile toxin (dmLT)23. Twenty-four hours

NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804

ac CpG + 2W-GFP dmLT + 2W-GFP 5 105 10 105 3.9 105 2.1

104 104 104 104

99.1 103 103 103 103 GFP

47.4 2 2 CD11c 10 10 102 102 0 0 0 0 0 102 103 104 105 0 102 103 104 105

102 103 104 105 102 103 104 105

0 0 GFP Isotype MAb W6 MAb CD11b 105 26.9105 27.2

104 104

103 103

b 102 102 Ovalbumin 2W-GFP fusion P=0.02 0 0 40 105 1.0 105 15 0 102 103 104 105 0 102 103 104 105

+ 30 104 104 W6 MAb

3 20 103 10 GFP 105 12.9 68.4105 11.5 75.8 2 102 10 10 % W6 MAb 0 0 104 104 0 0102 103 104 105 0102 103 104 105 OVA 2W-GFP 103 103 W6 MAb CD11c 102 102 0 0

0 102 103 104 105 0 102 103 104 105

CD11b

Figure 4 | Functional validation of W6 (2W:IAb) MAbs. (a) In vitro antigen presentation on bone marrow-derived dendritic cells isolated from C57BL/6 mice. Cells were stained with W6 MAb after 24 h pulse with GFP covalently linked to 2W peptide. GFP and W6 MAb double-positive cells are shown compared to isotype control (insert). Data are representative from two independent experiments with 3 mice per group. (b) In vivo immunostaining of 2W antigen-presenting cells. Mice were immunized intradermally in the ear with either Ovalbumin (OVA) or 2W-GFP. 24 h post immunization, cervical lymph nodes were collected, dissociated and gated for lymphocyte size, singlets, CD19 À and MHCII þ . Statistical significance was calculated using a two-tailed Student’s t-test. Data are representative of two independent experiments with four animals per group. (c) Mice were immunized with 2W-GFP and either CpG or dmLT23. Twenty-four hours later, draining lymph nodes were assayed using W6 MAb. Shown are percentages of CD19 À , CD11b þ CD11c þ dendritic cells. Data are representative from four independent experiments with 2–3 mice per group. later, draining lymph nodes were assayed by flow cytometry for diabetic recipients followed by injection of ethylene-carbodiimide antigen-specific presentation using the W6 antibody. Shown in (ECDI)-fixed p63 peptide-coupled cells (p63cc) to induce Fig. 4c, the W6 MAb identified 27% of the DCs containing GFP tolerance14,16 and either control or FS1 MAb and monitored compared with 3% in the isotype control group and B70% these the mice for diabetes. P63cc completely prevented diabetes cells were CD11b þ CD11c þ CD19 À dendritic cells. These results induction in 100% of the mice, while control bovine serum demonstrate that W6 MAb can identify different subsets of albumin-coupled cell-treated mice developed severe diabetes antigen-presenting cells in vivo. (Fig. 5b). Mice given p63cc and FS1 MAb developed diabetes, indicating the FS1 MAb prevented the induction of antigen- In vivo functional validation of p:MHCII MAb. Next, we sought specific tolerance in vivo. to determine whether MAbs directed against p:MHCII could The W6 MAb was next evaluated for its ability to block prevent TCR recognition in vivo to limit T-cell proliferation. 2W-specific in vivo expansion of antigen-specific T cells in NOD mice were challenged with p63 peptide plus lipopoly- response to an acute or chronic bacterial infection. C57BL/6 mice saccharide (LPS) with FS1 MAb or Y-Ae3 (anti-Ea:IAb)asa were administered W6 MAb and infected with Listeria negative control. Four days post challenge we measured a monocytogenes expressing 2W25. Seven days later the number significant reduction in antigen-specific T-cell expansion with of activated 2W-specific cells was decreased by 146-fold in FS1 MAb administration (Fig. 5a). Using dual fluorochrome response to W6 MAb compared with no antibody control tetramer staining and flow cytometry, we detected 30-fold (Fig. 6a). To test in vivo 2W-specific CD4 þ T-cell responses and expansion of p63 specific T cells stimulated with p63 þ their contribution to bacterial clearance in a chronic infection, LPS þ Y-Ae MAb control, compared with only a twofold wild type 129 S1 mice were infected with 108 Salmonella expansion with p63 þ LPS þ FS1 MAb over naive mice Typhimurium expressing 2W peptide26,27. The mice received a (Fig. 5a). In addition to decreased expansion, we also single dose of W6 blocking antibody 14 days post infection and determined that the FS1 MAb decreased activation and cell were sacrificed at day 35 to evaluate antigen-specific T-cell cycle progression (Fig. 5a). Next, we used the FS1 MAb in vivo to proliferation and colony forming units. Infected mice treated with prevent antigen-specific tolerance, resulting in rapid autoimmune the W6 MAb had significantly lower 2W-specific CD4 þ T cells, diabetes19,24. Specifically, we transferred activated BDC2.5 T cells higher bacterial burden, and did not clear the Salmonella after 4 days of in vitro stimulation into wild-type NOD pre- infection (Fig. 6b,c). These data highlight the critical

6 NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804 ARTICLE

a P=0.0128

105 0.3 0.03 105 P=0.0091 P=0.0018 7.1 40 104 104 Naive 30 103 103 0 20 0 Q4 6.2E-3 0 103 104 105 0103 104 105 10 Fold expansion relative to naive 5 105 0.5 0.87 10 85.0 0 104 104 P=0.0007 LPS+p63+Y-Ae 103 103 100 P<0.0001 P=0.0014

0 + 80 0 Q4 4.4E-3 Ki67

+ 60 0103 104 105 0103 104 105 40 105 0.1 0.08 105 59.2

%CD44 20 104 104

APC 0 LPS+p63+FS1 g7 103 103 0 Naive 0 3.7E-3 p63:IA CD44 0 103 104 105 0103 104 105 p63:IAg7 PE Ki67 LPS+p63+Y-AeLPS+p63 MAb +FS1 MAb b 100

80

60 BSAcc BSAcc + FS1 MAb 40 p63cc

% Diabetes free p63cc + FS1 MAb 20

0 0 5 10 15 Days post coupled cell transfer

Figure 5 | In vivo blockade of T-cell proliferation and prevention of T-cell tolerance following FS1 administration. (a) In vivo blockade of antigen-specific proliferation and cell cycle progression 4 days after i.v. administration of FS1 MAb plus p63 and LPS compared to Y-Ae (anti-IEa:IAb), control treatment or untreated naı¨ve mice. P63:IAg7 tetramer PE and APC double postive cells from the spleen were enriched and gated on lymphocyte size, singlets, live cells, B220 À , CD11c À , CD11b À , CD3e þ and CD4 þ . Statistical significance was calculated using a two-tailed Student’s t-test. Data are representative from two independent experiments with 2–3 mice per group. (b) In vivo blockade of antigen-specific tolerance using FS1 MAb to prevent ethylene-carbodiimide antigen-fixed-coupled cell tolerance. NOD mice received activated BDC2.5 TCR transgenic CD4 þ T cells to induce diabetes. Ten mice per group received either p63-coupled cells or BSA-coupled cells. Five mice per treatment received FS1 MAb or control. Mice were followed daily for diabetes after day 3 by blood glucose measurements. Data are representative from two independent experiments. importance of a single antigen-specific T-cell population and the prevent tolerance induction and bacterial pathogen clearance. blocking ability of the W6 MAb to prevent in vivo pathogen We have demonstrated that this novel methodology is highly clearance. efficient due to pre-screening and enrichment, saving both time To compare the affinity of the FS1 and W6 MAb with and resources (Table 1). This approach will be useful to generate previously published reagents, we performed a direct side-by-side additional p:MHCII-targeted MAbs as very few exist. The comparison with known IAg7 or IAb specific antibodies. most well-known is the Y-Ae antibody (anti-Ea:IAb), which b The results are shown in Table 2, and illustrate that the FS1 recognizes Ea52–68 bound to IA MHCII molecules, and was MAb (anti-p63:IAg7) has 100-fold higher affinity (1.7 Â 10 À 11) used to understand central tolerance and alloreactive antigen 28 g7 À 9 3,31,32 than the 10–2.16 MAb (anti-IA ) (2.9 Â 10 KD (M)). The presentation . The efficiency of Y-Ae generation has not W6 MAb (anti-2W:IAb) had an affinity comparable to known been described in the literature3, however, two more recent IAb antibodies (Y3P29, Y-Ae3 and AF6-120.1 (ref. 30)). These reports address this issue. The generation of MAbs specific for results suggest that the two MAb generated had comparable or two different peptides of hen egg lysozyme (HEL), specifically two k higher affinity than conventional approaches used to develop HEL46–61:IA (clones B6G and C4H) MAb were identified by MAb. screening 500 clones8. An additional clone specifically k recognizing HEL116–129:IA (D8H) was identified by screening Discussion 500 different colonies. However, these clones were found to also Using this methodology we have generated hybridomas produ- weakly stain cells expressing IAk in a HEL-independent manner8. cing six novel anti-peptide:MHCII MAb, and for two of these In a separate report, Dadaglio et al.9 generated a clone specific for k presented here, we demonstrate the high affinity for antigen and HEL48–62:IA (Aw3.18) by screening 1,000 colonies. Thus, the biological capacity to limit TCR engagement, prevent subsequent efficiency of generating a MAb clone using conventional T-cell activation, label antigen-presenting cells, and in vivo use to approaches has been reported to range from 1:250 to 1:1,000 or

NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804

a P=0.0249 Isotype W6 MAb P=0.0005 P=0.0026 250 105 34.2 105 1.68 200 104 104 150 103 103

CD44 100 102 102 0 0 50 Fold expansion relative to naive 0 0102 103 104 105 0102 103 104 105

Naive b Control 2W:IA -PE W6 MAb bc 4 106 P=0.04 10 P=0.0298 103 105 cells

b+ 102

4

10 CFU/organ 2W:I-A 101

103 100 Untreated W6 MAb Untreated W6 MAb

Figure 6 | In vivo blockade of T-cell proliferation and prevention of bacterial clearance following W6 administration. (a) C57BL/6 mice were infected with Listeria monocytogenes expressing 2W. Seven days post infection and W6 antibody treatment, 2W-specific cells were magnetically enriched from the spleen. Cells were gated on lymphocyte size, singlets, CD19 À , F4/80 À , CD11c À , CD11b À , CD3e þ and CD4 þ . Shown are representative FACS plots for CD44hi versus 2W:IAb tetramer PE þ cells and fold expansion of antigen-specific cells relative to uninfected mice. Data are representative of two independent experiments with 3–5 mice per group. (b,c) Wild type 129 S1 mice were infected with Salmonella Typhimurium expressing 2W, and 14 days following infection mice were treated with blocking W6 MAb. At day 35 post infection (21 days following MAb administration), (b), the number of 2W- specific T cells was determined from the spleen as described in a and (c) the colony forming units for the spleen was determined. Statistical significance was calculated using a two-tailed Student’s t-test. Data include 10 mice from two independent experiments.

immunization and tetramers for B-cell enrichment. With Table 2 | Comparison of anti-p:MHCII antibody affinities. advances of our understanding of peptide binding to both MHCI and MHCII, enhanced algorithms can be developed to better Antibody clone Antigen KD(M) predict peptide register binding leading to greater numbers of FS1 p63:IAg7 1.7 Â 10 À 11 6,34,35 g7 À 9 p:MHC monomers . This, combined with high throughput 10–2.16 (ref. 28) p63:IA 2.9 Â 10 generation of p:MHCII monomers by peptide exchange36 would W6 2W:IAb 3.4 10 À 9 Â alleviate the current limitations of this approach. The generation AF6-120.1 (ref. 30) 2W:IAb 7.1 Â 10 À 9 Y3P (ref. 29) 2W:IAb 1.9 Â 10 À 9 of MAb targeting mouse p:MHC class I (MHCI) could also be b À 10 Y-Ae (ref. 3) Ea52–68:IA 4.8 Â 10 performed using this methodology. In addition, human MAb targets against peptides in HLA class I or II would also be amenable to this approach, thus offering great potential for (0.1–0.4%). In Table 1, we report a range of human immunotherapy in the context of autoimmunity, tumour 1:18–1:115 using the novel methodology described here immunity or allograft rejection. The ability to specifically block (0.9–13.5%). Taken together, our success rate is 2.25 to 33.75- antigen presentation to T or B cells during bacterial or viral fold higher than traditional approaches resulting in a higher pathogenesis could provide mechanistic insight for immunity and efficiency. The fact that fewer clones are required for positive regulation, particularly with respect to persistent infection. identification also results in a significant advantage of lower costs Finally, the use of MAb specific for p:MHC allows the study of and fewer hours required for hybridoma screening. specific subsets of antigen-presenting cells during every aspect of More recently, an antibody against insulin B peptide9–23 in immune recognition ranging from immune homoeostasis to g7 g7 IA (anti-InsB9–23:IA ) was generated and shown to defining novel roles for multiple subsets of antigen-presenting delay diabetes in NOD mice4. In the current study, the FS1 cells responding to vaccination and infection. MAb was generated against a diabetes-relevant peptide in the We anticipate the need for immunotherapeutics in the form of context of IAg7. Here, we demonstrate successful blockade of MAbs directed against p:MHCII will increase as new pathways antigen-specific tolerance using FS1 MAb causing rapid diabetes and novel antigens are identified during disease. These targets (Fig. 5b). Future work using FS1 MAb will characterize the role could include not only foreign antigens, but also allogeneic for polyclonal or monoclonal T cells during spontaneous antigens during transplantation, tumour neo-antigens, autoimmune diabetes and tolerance16,33. The W6 reagent self-proteins targeted during autoimmunity, or bacterial and viral was generated to understand T-cell responses during both antigens from infected cells. As these targets are identified, there homoeostasis and bacterial pathogenesis, as the 2W peptide will be a great need to limit antigen-specific T-cell responses, and can be engineered into a pathogen of interest. Future work the methodology described here offers a more efficient approach with this reagent will characterize cells presenting antigen in over conventional protocols. response to bacterial infections or immunization with different adjuvants. Methods One limitation to this approach centers on possessing knowl- Mice. Female NOD mice (6–8 weeks of age) were purchased from Taconic. Female edge of the antigen target, availability of p:MHC monomers for C57BL/6 (6–8 weeks of age), female 129 S1 (6–8 weeks of age) and female BALB/c

8 NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804 ARTICLE mice (8–10 weeks of age) were purchased from the Jackson Laboratory. Female boosted by intravenous (i.v.) injection of 100 ml total volume containing 25 mgof (6-week-old) NOD.BDC2.5 Thy1.1 transgenic mice were bred under specific p:MHCII in PBS. Three days after boost a single cell suspension was made from pathogen-free, barrier facility at the University of Minnesota. Animals were housed pooled spleens and draining lymph nodes. Cells were stained in 150 ml of complete under specific pathogen-free, barrier facility in accordance with NIH guidelines. media (DMEM, 10% FCS, b-ME, pen/strep, nonessential amino acids) containing All animal procedures were approved by the University of Minnesota or Tulane 13.3 nM PE-conjugated p:MHCII/mouse and incubated on ice for 25 min. Cells Institutional Animal Care and Use Committee. were washed with complete media and resuspended in 200 ml of complete media containing 50 ml of anti-PE microbeads (Miltenyi Biotec)/mouse and incubated on Peptides. Peptides used for in vivo immunization and in vitro stimulation, ice for 25 min. Cells were washed with complete media and resuspended in 5 ml of and peptide pulsing include p63 (RTRPLWVRME), p31 (YVRPLWVRME), complete media/spleen. The cell suspension was then applied to a pre-equilibrated LS magnetic column (Miltenyi Biotec) and washed 2 Â with 3 ml of complete OVA141–160 (CARELINSWVESQTNGIIRN) (Genemed Synthesis), 2W (EAWGALANWAVDSA) (Genscript). media. The cells were eluted from the column in 5 ml medium A (Stem Cell Technologies), centrifuged and enumerated. The enriched B cells were fused with SP2/0 mouse myeloma cells using the HY Hybridoma Cloning Kit according the Peptide:MHCII monomers and tetramers. p63:IAg7 (RTRPLWVRME), g7 b manufacturer’s protocol using method A (Stem Cell Technologies). A small portion InsB10–23:IA (HLVERLYLVCGEEG), 2W:IA (EAWGALANWAVDSA) and b of the enriched cells and flowthrough was stained as described above to determine LLO:IA (NEKYAQAYPNVS) was either from the NIH tetramer core facility antigen-specific B-cell purity and phenotype before hybridoma fusion. (Emory University) or produced using the S2 insect-cell expression system12–14. Briefly, peptide:IAg7 or peptide:IAb molecules were expressed in Drosophila S2 cells using the DES Drosophila Expression System kit (Invitrogen). The S2 cells Hybridoma selection and specificity screening. Twelve days post-HAT were co-transfected using calcium phosphate, with plasmids encoding the alpha selection, individual colonies were handpicked and transferred to 96-well plates chain, the peptide-linked beta chain, BirA ligase and a blasticidin resistance gene at containing medium E (Stem cell technologies). Four days later hybridoma a molar ratio of 9:9:9:1 for IAb and IAg7. Transfected cells were selected in supernatants were transferred to 96-well plates and fresh medium E was added to blasticidin-containing Schneider’s Drosophila medium (Invitrogen) with 10% fetal the cells. To test the specificity of hybridoma supernatants for MHCII and À 1 m À 1 p:MHCII, a decoy screening approach was employed. ELISA plates were coated bovine serum (FBS), 100 U ml penicillin/streptomycin (Gibco) and 20 gml À 1 g7 g7 g7 with 50 ng well either p63:IA or InsB10–23:IA (for p63:IA antibodies), or gentamycin (Invitrogen) for 1 week at 28 °C, passaged into serum-free media b b containing 25 mgmlÀ 1 blasticidin (Invivogen), and scaled to 0.5 l cultures in 2 l 2W:IA or LLO:IA , and then blocked with 1% BSA in PBS for 1 h. Hybridoma shaker flasks maintained at 150 r.p.m. When cell densities reached 5 Â 106 ml À 1, supernatants were mixed 1:1 with ELISA wash buffer (PBS þ 0.05% Tween20) and monomer expression was induced by the addition of 0.8 mM copper sulphate. added to the p:MHCII-coated plates and incubated at 37 °C for 2 h. Media alone b g7 was used as a negative control while anti-IAg7 (clone 10–2.16 (ref. 28), BioXcell) or Peptide:IA or IA heterodimers were purified from supernatants 8 days later by b 29 immobilized metal ion affinity chromatography using a His-Bind purification kit anti-IA (Y3P , ATCC) were used as a positive control. For antibody detection the (Novagen-EMD Biosciences) and eluted using 1M imidazole. The biotinylated wells were incubated with HRP-conjugated goat anti-mouse IgG (BioLegend) pMHCII heterodimers in the eluate were then affinity purified using Monomeric diluted to 1:2,000 at room temperature for 2 h followed by addition of ABTS Avidin UltraLink (Pierce). Bound peptide:IAb or IAg7 molecules were eluted with substrate solution (KPL) and detection by absorbance at 405 nm. Antibodies reacting to both p:MHCII monomers were considered specific for MHCII 2 mM biotin in PBS and excess-free biotin was removed by centrifugation and four g7 b washes with 12 ml PBS using a 30 KD cut-off Amicon Ultra-15 filter (Millipore). independent of peptide, while antibodies reacting with only p63:IA or 2W:IA Tetramers were produced by incubating p:MHCII monomers with streptavidin- were considered p:MHCII specific. APC (Prozyme # PJ27S) or streptavidin-PE (Prozyme #PJRS27) at a 4:1 molar ratio. For the detection of SA and fluorochrome specific B cells, AF647 was Antibody affinity measurements and sequencing. The affinity of the two novel conjugated to SA-PE (Prozyme) for 60 min at room temperature using an antibody MAb generated (FS1 and W6) were directly compared with known IAg7 or IAb labelling kit (ThermoFisher Scientific) and free AF647 was removed by specific antibodies (10–2.16 (ref. 28) (BioXcell), Y3P29 (ATCC), Y-Ae3 centrifugation in a 30 KD molecular weight cut-off filter. The concentration was (eBioscience) and AF6-120.1 (ref. 30) (BioLegend)) by Bio-Layer Interferometry by then adjusted to 1 mM PE based on the absorbance at 565 nm using a nandrop Precision Antibody (Columbia, MD). Antibody sequences were obtained using the spectrophotometer (ThermoFisher Scientific). Similarly, SA-APC (Prozyme) was SMARTer RACE cDNA Amplification Kit (Clontech) according to the conjugated to DyLight 755 using an antibody labelting kit (ThermoFisher manufacturer’s instructions. Primers used for reverse transcription were Scientific) and the concentration was adjusted to 1 mM APC based on the GATTACGCCAAGCTTTATGCAAGGCTTACAACCACA (heavy chain), absorbance at 650 nm. GATTACGCCAAGCTTCACAATTTTCTTGTCCACCTTGGTGC (heavy chain nested), GATTACGCCAAGCTTCTCATTCCTGTTGAAGCTCTTGACAAT Antigen-specific B-cell enrichment and phenotyping. BALB/c mice were (kappa light chain), GATTACGCCAAGCTTACACTCAGCACGGGACAAA immunized with 50 mg of pMHCII emulsified in complete Freund’s adjuvant CTCTTCTC (lambda light chain 1, 4), GATTACGCCAAGCTTACACTCTGCAG (CFA, Sigma) subcutaneously in the base of the tail. Seven days post immunization GAGACAGACTCTTTTC (lambda 2,3). single-cell suspensions from spleen and pooled lymph nodes (inguinal, brachial, cervical and axillary) were prepared by forcing the tissue through a 100 mm cell Proliferation and cytokine production. Cells were isolated from spleen and strainer using the plunger end of a 1 ml syringe, washed with RPMI containing 2% peripheral lymph nodes of NOD.BDC2.5 transgenic mice, the red blood cells were FBS and resuspended in 100 ml Fc block (2.4G2, 0.05% sodium azide). The cells lysed by Tris-buffered ammonium chloride, and the cells were labelled with CFSE were next incubated with 5 nM SA-PE-AF647, and SA-APC-DyLight755 for (ThermoFisher Scientific). Labelled cells were resuspended in complete DMEM 10 min at 25 °C, followed by peptide:MHCII-conjugated PE and APC tetramers at media at a final concentration of 4 Â 106 cells ml À 1 and p63 or p31 peptide was 10 nM for 25 min on ice in a final staining volume of 200 ml. The cells were then added to a final concentration of 0.05 mM. Purified monoclonal antibody or 50 mlof washed with 12 ml PBS þ 2% FBS, resuspended in 150 ml of PBS þ 2% FBS, mixed hybridoma supernatant was added to each well containing 200 ml of cells in a with 25 ml anti-PE and anti-APC MicroBeads (Miltenyi Biotec) and incubated for 96-well plate, and incubated for 4 days at 37 °C with 5% CO2. Cells were collected, 25 min at 4 °C. The cells were then washed with 12 ml PBS þ 2% FBS, resuspended resuspended in 2.4G2 Fc block for 10 min at 4 °C, and stained at 1:100 dilution for in 3 ml PBS þ 2% FBS and applied to a magnetized LS column (Miltenyi Biotec). 30 min at 4 °C with antibodies against CD4-BV510 (RM4-5, BD Biosciences), The column was washed three times with 3 ml PBS þ 2% FBS, removed from the CD8a-BV650 (53–6.7, BD Biosciences), CD3e-PerCp-Cy5.5 (KT4, BD Bios- magnet and the cells were eluted in 5 ml PBS þ 2% FBS. All the eluted cells and ciences), B220-ef450 (RA3-6B2, eBioscience), CD11c-ef450 (N418, eBioscience), 1/20th of the flowthrough were then centrifuged and stained with surface CD11b-ef450 (M1/70, eBioscience), TCR Vb4-PE (KT4, BD Biosciences) and dead antibodies IgM-BV421 (RMM-1, BioLegend), B220-PE-Cy7 (RA3-6B2, Tonbo cells were gated out using a viability ghost red dye (Tonbo), and run on a LSRII Biosciences), CD38-AF700 (90, eBioscience), GL7-FITC (GL7, eBioscience), Fortessa X-20 flow cytometer (Becton Dickinson) and analyzed using FlowJo IgD-BV786 (11-26 c.2a, BD Biosciences), CD11b-BV510 (M1/70, BD Biosciences), software (v10). For antibody dose–response curves, cells were isolated from CD11c-BV510 (N418, BioLegend), F4/80-BV510 (BM8, BioLegend), CD90.2- NOD.BDC2.5 transgenic mice, labelled with CFSE as described above and cultured BV510 (53-2.1, BioLegend), Live/Dead Ghost 510 dye (Tonbo Biosciences) for with p31 or p63 (0.05 mM) with varying doses of FS1 MAb or isotype control 30 min at 4 °C. Next, the cells were fixed with Cytofix/Cytoperm solution (BioXcell). After a 4-day incubation at 37 °C with 5% CO2, cells were collected for (BD Biosciences) for 20 min at 4 °C, washed twice with permeabilization buffer and the flow cytometry, while supernatants were analyzed using ProcartaPlex Assay kit stained with intracellular antibodies IgG (H þ L)-AF350 (polyclonal, ThermoFisher EPX170-26087-901 (eBioscience). Scientific) for 30 min at 4 °C in permabilization buffer before the flow cytometry analysis on a LSRII Fortessa instrument (Becton Dickinson) equipped with five Antibody staining of peptide-pulsed antigen-presenting cells. Purified laser lines (355, 405, 488, 561 and 640 nm). All antibodies were used at a 1:100 monoclonal antibody was directly conjugated to Alexa Fluor-488 by protein dilution for staining, except CD90.2, which was diluted 1:500. The data were labelling kit (ThermoFisher Scientific) according to the manufacturer’s protocol. analyzed using FlowJo software (v.10). Splenocytes were isolated from NOD mice, red blood cells lysed and resuspended 6 À 1 at in complete DMEM at 6 Â 10 cells ml . p63 or OVA141–160 was added to the Isolation and fusion of antibody-producing B cells. BALB/c mice were media at a final concentration of 40 mM and the cells were incubated for 1.5 h at immunized subcutaneously in the base of the tail with 50 mg of p:MHCII monomer 37 °C with 5% CO2. Cells were collected incubated in 2.4G2 for 10 min on ice and emulsified in complete Freunds’ adjuvant. Twenty eight days later each mouse was stained at 1:100 dilution for 30 min at 4 °C with FS1-AF488 and antibodies against

NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804

CD4-BUV395 (GK1.5, BD Biosciences), CD8a-APC-ef780 (53–6.7, eBioscience), References CD3e-BV650 (145-2C11, BD Biosciences), B220-PE (RA3-6B, Tonbo), 1. Kohler, G. & Milstein, C. Continuous cultures of fused cells secreting antibody CD11c-PE-Cy (N418, eBioscience), CD11b-PerCp-Cy5.5 (M1/70, Tonbo), of predefined specificity. Nature 256, 495–497 (1975). F4/80-APC (BM8.1, Tonbo), IAg7-biotin (10.2-16, BioXcell), SA-BV421 2. Yokoyama, W. M. et al. Production of monoclonal antibodies. Curr. Protoc. (BioLegend) and dead cells were gated out using Ghost violet 510 viability dye Immunol. 102, Unit 2.5 (2013). (Tonbo), and run on a LSRII Fortessa X-20 flow cytometer (Becton Dickinson) and 3. Murphy, D. B. et al. A novel MHC class II epitope expressed in thymic medulla analyzed using FlowJo software (v10). but not cortex. Nature 338, 765–768 (1989). 4. Zhang, L. et al. Monoclonal antibody blocking the recognition of an insulin In vivo blockade of T-cell activation and tolerance. The NOD mice were peptide–MHC complex modulates type 1 diabetes. Proc. Natl Acad. Sci. USA administered 50 mg acetylated-p63 peptide with either 250 mg FS1 or Y-Ae MAb in 111, 2656–2661 (2014). PBS containing 2 mg LPS (Sigma) by i.v. injection in the tail vein. Four days post 5. Wolpl, A. et al. Human monoclonal antibody with T-cell-like specificity injection splenocytes were isolated and red blood cells lysed as above. Cells were recognizes MHC class I self-peptide presented by HLA-DR1 on activated cells. g7 stained with both APC and PE-conjugated p63:IA tetramers followed by Tissue Antigens 51, 258–269 (1998). 14,37 magnetic enrichment of double tetramer positive cells as previously described . 6. Muraille, E. et al. Direct visualization of peptide/MHC complexes at the surface Enriched cells were incubated with 2.4G2 Fc block for 10 min at 4 °C and stained and in the intracellular compartments of cells infected in vivo by Leishmania ° with surface antibodies at 1:100 dilution for 30 min at 4 C against CD3e-APC- major. PLoS Pathog. 6, e1001154 (2010). ef780 (17A2, eBioscience), CD4-BUV395 (GK1.5, BD Biosciences), CD8a-BV650 7. Baldwin, K. K., Reay, P. A., Wu, L., Farr, A. & Davis, M. M. A T-cell (53–6.7, BD Biosciences), CD44-BV786 (IM7, BD Biosciences), B220-PerCp-Cy5.5 receptor-specific blockade of positive selection. J. Exp. Med. 189, 13–24 (1999). (RA3-6B2, eBioscience), CD11b-PerCp-Cy5.5 (M1/70, Tonbo), CD11c-PerCp- 8. Zhong, G., Reis e Sousa, C. & Germain, R. N. Production, specificity, and Cy5.5 (N418, eBioscience) and Ghost violet 510 viability dye (Tonbo) for 30 min at 4 °C. The cells were then fixed using the Foxp3 staining buffer kit (eBioscience) and functionality of monoclonal antibodies to specific peptide-major stained with anti-Ki67-PE-Cy7 (SolA15, eBioscience) antibodies for 1 h at 4 °C, and histocompatibility complex class II complexes formed by processing of run on a LSRII Fortessa X-20 flow cytometer (Becton Dickinson) and analyzed exogenous protein. Proc. Natl Acad. Sci. USA 94, 13856–13861 (1997). using FlowJo software (v10). Pre-diabetic NOD mice received in vitro activated 9. Dadaglio, G., Nelson, C. A., Deck, M. B., Petzold, S. J. & Unanue, E. R. BDC2.5 T cells (ref. 19) followed by 30 million p63-cc intravenously and 500 Characterization and quantitation of peptide–MHC complexes produced micrograms of FS1 antibody on day 0 and 2 intravenously. Mice were monitored from hen egg lysozyme using a monoclonal antibody. Immunity 6, 727–738 for diabetes by daily blood glucose measurements. (1997). 10. Pape, K. A., Taylor, J. J., Maul, R. W., Gearhart, P. J. & Jenkins, M. K. Different Infections. C57BL/6 mice were injected intravenously with 107 actA-deficient B cell populations mediate early and late memory during an endogenous Listeria monocytogenes expressing 2W protein25, and on the same day injected immune response. Science 331, 1203–1207 (2011). intravenously with 500 mg of W6 (anti-2W:IAb) blocking antibody. Seven days 11. Newman, J., Rice, J. S., Wang, C., Harris, S. L. & Diamond, B. Identification of later, splenocytes were isolated and stained as above and magnetically enriched for an antigen-specific B cell population. J. Immunol. Methods 272, 177–187 2W:IAb -PE tetramer. Wild type 129S1 mice were infected with Salmonella (2003). Typhimurium expressing 2W, and 14 days following infection mice were treated 12. Moon, J. J. et al. Tracking epitope-specific T cells. Nat. Protoc. 4, 565–581 intravenously with 500 mg of blocking W6 MAb. Thirty-five days post infection, the (2009). spleen was removed and made into a single cell suspension and 10% was plated for 13. Moon, J. J. et al. Naive CD4( þ ) T cell frequency varies for different bacterial CFU with the addition of 0.1% Triton (ref. 27). The remaining cell epitopes and predicts repertoire diversity and response magnitude. Immunity suspension was used for enrichment of 2W specific cells as described above. Cells 27, 203–213 (2007). were run on a LSRII Fortessa X-20 flow cytometer (Becton Dickinson) and 14. Pauken, K. E. et al. Cutting edge: type 1 diabetes occurs despite robust anergy analyzed using FlowJo software (v10). among endogenous insulin-specific CD4 T cells in NOD mice. J. Immunol. 191, 4913–4917 (2013). Footpad immunization, antigen-presenting cell isolation and staining. The 15. Fife, B. T. & Bluestone, J. A. Control of peripheral T-cell tolerance and NOD hind limb footpads were injected with 100 ml total volume containing 200 mg autoimmunity via the CTLA-4 and PD-1 pathways. Immunol. Rev. 224, of p63 or OVA141–160 peptide in PBS. One and half hours later the popliteal lymph 166–182 (2008). nodes were removed, minced and digested in RPMI containing 2% FCS, 16. Pauken, K. E., Jenkins, M. K., Azuma, M. & Fife, B. T. PD-1, but not PD-L1, À 1 À 1 collagenase D (40 U ml ) and DNase I (250 mgml ) for 30 min at 4 °C. Cells expressed by islet-reactive CD4 þ T cells suppresses infiltration of the pancreas were then washed with Hanks balanced salt solution containing 5 mM EDTA and during type 1 diabetes. Diabetes 62, 2859–2869 (2013). 2% FCS, centrifuged and stained with surface antibodies and analyzed by flow 17. Haskins, K., Portas, M., Bradley, B., Wegmann, D. & Lafferty, K. T-lymphocyte cytometry as described above using the FS1-AF488 antibody. clone specific for pancreatic islet antigen. Diabetes 37, 1444–1448 (1988). 18. Katz, J. D., Wang, B., Haskins, K., Benoist, C. & Mathis, D. Following a Ear pinna immunization, antigen-presenting cell isolation and staining. diabetogenic T cell from genesis through pathogenesis. Cell 74, 1089–1100 C57BL/6 mice were immunized intradermally in the ear pinna with 10 mg either (1993). 23 Ovalbumin (OVA) or 2W-GFP plus 1 mg dmLT (a gift from J. Clements) or 10 mg 19. Fife, B. T. et al. Interactions between PD-1 and PD-L1 promote tolerance CpG (Sigma). After 24 h, the cervical lymph nodes were removed, dissociated using À 1 by blocking the TCR-induced stop signal. Nat. Immunol. 10, 1185–1192 (2009). a 100-mm mesh and mechanically disrupted, and digested with 300 Mandl U ml 20. Taylor, J. J. et al. Deletion and anergy of polyclonal B cells specific Collagenase D (Roche Applied Sciences) for 30 min at 37 °Cin1Â PBS þ 2% FBS. for ubiquitous membrane-bound self-antigen. J. Exp. Med. 209, 2065–2077 Cells were then washed 1 Â PBS þ 2% FBS, centrifuged and Fc receptors were (2012). blocked in 100 ml 2.4G2 hybridoma supernatant containing 2% rat and mouse serum for 10 min at room temperature. For surface staining, 1 mg of biotinylated 21. Judkowski, V. et al. Identification of MHC class II-restricted peptide ligands, W6 antibody was added to the cells and incubated on ice for 45 min, washed with including a glutamic acid decarboxylase 65 sequence, that stimulate PBS þ 2% FCS, followed by staining for 30 min at 4 °C with antibodies diabetogenic T cells from transgenic BDC2.5 nonobese diabetic mice. CD11c-PerCp-Cy5.5 (N418, BioLegend), CD11b-AF700 (M1/70, BioLegend), J. Immunol. 166, 908–917 (2001). CD19-ef450 (1D3, eBioscience), MHCII-FITC (M5/114.15.2, BioLegend), 22. Rees, W. et al. An inverse relationship between T cell receptor affinity and DEC-205-PE (205yekta, eBioscience) and streptavidin-APC (BioLegend). All antigen dose during CD4( þ ) T cell responses in vivo and in vitro. Proc. Natl antibodies were used at 1:100 dilution except CD11c and CD19 at 1:40 and CD11b Acad. Sci. USA 96, 9781–9786 (1999). at 1:200. The cells were then run on a LSRII flow cytometer (Becton Dickinson) 23. Norton, E. B. et al. The novel adjuvant dmLT promotes dose sparing, mucosal and analyzed using FlowJo software (v10). immunity and longevity of antibody responses to the inactivated polio vaccine in a murine model. Vaccine 33, 1909–1915 (2015). Statistical analysis. Data display and statistical analysis was conducted using 24. Fife, B. T. et al. Insulin-induced remission in new-onset NOD mice is Prism software (GraphPad Prism v6). Statistical significance was analyzed using the maintained by the PD-1-PD-L1 pathway. J. Exp. Med. 203, 2737–2747 (2006). two-tailed Student’s t-test for comparison of two mean. Values of Pr0.05 were 25. Ertelt, J. M. et al. Selective priming and expansion of antigen-specific considered significant. Foxp3-CD4 þ T cells during Listeria monocytogenes infection. J. Immunol. 182, 3032–3038 (2009). Data availability. Sequence data that support the findings of this study have been 26. Uzzau, S., Figueroa-Bossi, N., Rubino, S. & Bossi, L. Epitope tagging of deposited in GenBank with the primary accession codes: FS1 heavy chain chromosomal genes in Salmonella. Proc. Natl Acad. Sci. USA 98, 15264–15269 KU955585 (www.ncbi.nlm.nih.gov/nuccore/KU955585), FS1 light chain KU955586 (2001). (www.ncbi.nlm.nih.gov/nuccore/KU955586), W6 heavy chain KU955587 27. Nelson, R. W., McLachlan, J. B., Kurtz, J. R. & Jenkins, M. K. CD4 þ T cell (www.ncbi.nlm.nih.gov/nuccore/KU955587) and W6 light chain KU955588 persistence and function after infection are maintained by low-level (www.ncbi.nlm.nih.gov/nuccore/KU955588). peptide:MHC class II presentation. J. Immunol. 190, 2828–2834 (2013).

10 NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications NATURE COMMUNICATIONS | DOI: 10.1038/ncomms11804 ARTICLE

28. Oi, V. T., Jones, P. P., Goding, J. W., Herzenberg, L. A. & Herzenberg, L. A. Program of the Cancer Research Institute (J.J.T.), Juvenile Diabetes Research Foundation Properties of monoclonal antibodies to mouse Ig allotypes, H-2, and Ia 2-2011-662 (B.T.F.) and 3-2014-215 (J.A.S.). antigens. Curr. Top. Microbiol. Immunol. 81, 115–120 (1978). 29. Janeway, Jr. C. A. et al. Monoclonal antibodies specific for Ia glycoproteins Author contributions raised by immunization with activated T cells: possible role of T cellbound Ia J.J.T. and M.K.J. designed and generated the 2W:IAb MAb. J.A.S., J.L.R. and B.T.F. antigens as targets of immunoregulatory T cells. J. Immunol. 132, 662–667 designed and generated the FS1 MAb (p63:IAg7). J.A.S., J.R.H., K.C.O., T.M. and B.T.F. (1984). designed, performed and analyzed the data to validate the p63:IAg7 MAb. J.J.T., D.I.K., 30. Lorber, M. I., Loken, M. R., Stall, A. M. & Fitch, F. W. I-A antigens on cloned M.K.J., D.R.F., B.J.R., S.J.L. and J.B.M. designed, performed and analyzed the data to alloreactive murine T lymphocytes are acquired passively. J. Immunol. 128, validate the W6 MAb (2W:IAb). J.A.S., J.J.T., M.K.J., J.B.M. and B.T.F. directed the 2798–2803 (1982). studies. J.A.S., D.R.F., J.B.M. and B.T.F. prepared the figures. J.A.S. and B.T.F. wrote the 31. Rudensky, A., Rath, S., Preston-Hurlburt, P., Murphy, D. B. & Janeway, Jr. C. A. manuscript. On the complexity of self. Nature 353, 660–662 (1991). 32. Murphy, D. B. et al. Monoclonal antibody detection of a major self peptide. MHC class II complex. J. Immunol. 148, 3483–3491 (1992). Additional information 33. Pauken, K. E. et al. Cutting edge: identification of autoreactive CD4 þ and Competing financial interests: The authors declare no competing financial interests. CD8 þ T cell subsets resistant to PD-1 pathway blockade. J. Immunol. 194, 3551–3555 (2015). 34. Lee, S. J. et al. Temporal expression of bacterial proteins instructs host CD4 T Reprints and permission information is available online at http://npg.nature.com/ cell expansion and Th17 development. PLoS Pathog. 8, e1002499 (2012). reprintsandpermissions/ 35. Nelson, R. W. et al. T cell receptor cross-reactivity between similar foreign and self peptides influences naive cell population size and autoimmunity. Immunity How to cite this article: Spanier, J. A. et al. Efficient generation of monoclonal antibodies 42, 95–107 (2015). against peptide in the context of MHCII using magnetic enrichment. Nat. Commun. 36. Day, C. L. et al. Ex vivo analysis of human memory CD4 T cells specific for 7:11804 doi: 10.1038/ncomms11804 (2016). hepatitis C virus using MHC class II tetramers. J. Clin. Invest. 112, 831–842 (2003). 37. Tubo, N. J. et al. Single naive CD4( þ ) T cells from a diverse repertoire produce This work is licensed under a Creative Commons Attribution 4.0 different effector cell types during infection. Cell 153, 785–796 (2013). International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise Acknowledgements in the credit line; if the material is not included under the Creative Commons license, This work was supported by NIH R01 AI106791 (B.T.F.), NIH U24-AI118635 (B.T.F.), users will need to obtain permission from the license holder to reproduce the material. NIH P01 AI35296 (B.T.F. and M.K.J.), NIH R01 AI103343 (J.B.M.), Irvington Fellowship To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

NATURE COMMUNICATIONS | 7:11804 | DOI: 10.1038/ncomms11804 | www.nature.com/naturecommunications 11